Noise And Dynamic Range

Two terms that are very important in assessing the quality of NMR data are signal to noise and dynamic range. The first has already been discussed, but it is useful to think about where noise comes from and what can be done to reduce it. The signal is proportional to the concentration of the sample and to the number of scans. Although signal is often measured as peak height, we must remember that it is the peak area that is proportional to concentration and number of scans; the peak height is very sensitive to shimming. As the shimming improves, the peak gets taller because the constant peak area is squeezed into a narrower and narrower peak. It is the peak height that determines whether a peak can be seen "over" the noise: If a peak is very broad, it is much easier to "lose" it in the noise than a sharp peak that "sticks up" above the level of the noise. If we compare two one-proton signals from the same molecule, the one with fewer splittings (J couplings) will be easier to see above the noise. For example, a fully resolved doublet of doublets (four peaks of equal height) is only one fourth of the height of a singlet with the same linewidth and area. Especially in low-sensitivity 2D experiments, the complex multiplet signals are often lost in the noise while the sharp singlet signals are easy to see. When measuring signal-to-noise ratio with a standard sample (0.1% Ethylbenzene in CDCl3 is the XH standard), the result is not meaningful unless the shimming is superb. With high-order shim errors, all of those "porches" and "pedestals" that may not even show above the noise are robbing intensity from the peak and lowering its peak height.

0 0

Post a comment